sashken EA - 3rd Place at Automated Trading Championship!

 

Thanks to sashken for sharing this winning EA. You can download it below.

The system is very simple:

Buy when MA 7 crosses above MA 20, and sell when MA 7 crosses below MA 20. Set TP to 140 and SL to 100.

You can follow the EA at ATC2006.

Enjoy!

Files:
sashken.mq4  6 kb
 

Ahhh a Moving Average expert? Didn't we prove that this didn't work well during choppy markets this fails? or at least it doesnt work well?

 

It'll fail in long term, unless you made change to the settings. I guess sashken optimized the EA to trade for the next 3 months, just like Hendrick did.

Ok, here's what I've observed today. The market is full of cycle that lasts one to three months. In each cycle, price moves in the same pattern every day. Ok, the winning systems have advantages trading in current cycle which seems to have started since Aug 1st. When this cycle ends, it will loss.

Look like the holygrail in forex is to be able to identify the cycle. I suspect that cycle is created by either GDP or Central Bank's Policy. I have the economic data releases on hand now, so i'll be testing this theory; then apply it on EvoX. If EvoX can avoid bad cycle, we'll finally beat the market at anytime we want to.

 

Observed a bit. The cycles have to do with USA FOMC Interest Rate Statement that is released monthly. Cycle in April 1st was created on the day FOMC changed the interest rate. And, the cycle in Aug was created when FOMC didn't change the interest rate after the meeting on Aug 8.

EvoX seems not working well when interest rate stays the same. And it double the account any month interest rate was changed. Very interesting! even more interesting when technical indicators don't tell nothing. TA can't detect cycle.

I'm going to apply this theory!

 

Hi ,You made a statement in a previous post that I was wondering if you could define a few of the terms:"In each cycle, price moves in the same pattern every day."Based on subsequent posts, the cycle refered to appears to be a broad market characteristic. You've indicated that the the cycle may be driven by the USA FOMC Interest Rate Statement that is released monthly. But what defines the cycle, the news release evet or whether there is an interest rate change or is there some othere characteristic?Also, please explain "price moves in the same pattern every day". What are the distinquishing characteristics of the pattern you've identified?_____Been a member for a while and this is my first post. If these questions have been answered elsewhere within the form, please let me know. Thanks

 

Ok, let's talk about breakout pattern from August till now. When the price breaks out 32 pips away from the resistance or support level, it'll retract back into the previous support and resistance. This pattern repeats every day from August. In big picture, you see the price creates only short-term trends lasting a day or two; then it goes back to moving sideway. You seldom see double breakout. In this cycle, you'll have to optimize your system to catch short trends instead. For instance, you lower from MA(14) to MA(7), MA(35) to MA(20), or 32 pips breakout to 8 pips. You'll have to go back to an optimizer program and bruteforce for the best settings. Metatrader has built-in optimizer, so if you're using EA, it's the best available tool to adapt your system. (Check Phoenix EA thread for optimization guide)

Before August, if price broke out 32 pips away from resistance or support, it would go out at least 33 pips more. You would see most trends were lasting weeks or the whole month. So this cycle is best used for breakout or trend following system.

It's not clear that interest rate switches the cycle, because there're many indicators released around it. It could be weekly jobless claims, or quarterly GDP, or even Fed Chairman/President's speaking.

The reason i'm interested in interest rate because central banks the dominant price movers in forex are only interested in such statement. Weekly jobless claim or GDP, why should they care? It's not their job. What they need to do is follow the blueprint in interest rate statement from FOMC meeting. These banks has to control the price range, the so-called interventions to make their currency stays close to specified price.

 

Hello,

Im new here in this forum and would first like to thank you for opening this great forum and for the wealth of information you offer.

Anyways ive been trying out the different EAs that you have posted, such as Phoenix and this one sashken, I was able to backtest them both and now im trying to forward test them, they have been running the whole day, phoenix alreadyp laced a sell position in USDJPY but sashken is still not doing anything... can someone give me a clue as to why sashken isint trading? its on the M15 chart...

thanks again,

Mikhail

 

Hi everybody,

My first post starts here

So according to you Scorpion (and so many other testers), EA relyiing on MAs will fail in the long term if its not constantly adapted to the market cycles. Then, let's say we want to keep the EA updated, on which periods should we do it? And how?

Thanks for your coming reply

 

Ok, sashken EA is a goner. One big loss made him go to the bottom of the list. The cycle has changed from now, and for months to come, this EA will fail in this new cycle. EURUSD hit 1.3000 yesterday, so i guess this is going to be a unique cycle.

 

Hi all,

Cycles do exist but how we measure or detect them techincally and can apply to trading pratically and sucessfully

 

When your EA fails many times consecutively, it means you need to optimize the settings for the new cycle. If you can't optimize, you need to implement the MM suitable for your system to survive the bad cycle.

I'm interested in Elder's 2% and 6% MM. He rules that you should risk only 2% of equity at any given time. Anytime your account loses 6% of equity from the end of last month, stop trading for the whole month. After that, you start with lowest lot, increase bit by bit. Until your account equity increases by 6% should you run it at full throttle. With this MM, you should minimize the loss during bad cycle and maximize the winning during good one.

The strategy seems awkward in forex, because 2% risk is too boring to trade. With 100:1 leverage and $10,000 equity, you can risk only 20 pips at any given time. When your account down 60 pips in November, you'd stop trading immediately and start again with 0.1 lot on December 1. With $1,000 equity, you should go with mini-lot. You then lose only 2% of equity for 20 pips, rather than 2 pips.

I'm going to test this strategy with my systems. If it works [in backtesting and forward testing], i'll publish my finding. And of course, integrate this MM in the new ZeroCode.